– Written by Juan Gabriel Covas 2021

This is my guide to compile OpenALPR under Linux CentOS 7 / RedHat, which was a pain since I didn't find any other similar how-to: the official page about OpenALPR just talks about Ubuntu.

OpenALPR is an open source Automatic License Plate Recognition library written in C++ […] that analyzes images and video streams to identify license plates. The output is the text of any license plate characters found.

Under Windows you can easily test OpenALPR using the official OpenALPR windows installers, but we want binaries for Linux CentOS…

The “rabbit hole” reference:

• OpenALPR github wiki about OpenALPR compilation instructions for Ubuntu Linux - The Easy Way

If you're feeling lucky, you could put all of this on a bash shell script and execute it, to see if you end up having the alpr command available to use.

# Installing OpenALPR from sources under CentOS 7

# some "easy" pre-requisites to be able to compile OpenALPR
sudo yum -y install epel-release
sudo yum -y install git wget
sudo yum -y groupinstall "Development tools"
sudo yum -y install cmake cmake3 curl curl-devel log4cplus-devel

# get OpenALPR sources
mkdir ~/src; cd ~/src 
git clone https://github.com/openalpr/openalpr.git
cd ~/src/openalpr/src && mkdir build; cd build

# compile Tesseract (and Leptonica), required by OpenALPR
sudo yum -y install autoconf-archive libjpeg-turbo-devel libpng-devel libtiff-devel zlib-devel
cd ~/src
wget http://www.leptonica.org/source/leptonica-1.81.1.tar.gz
tar -zxvf leptonica-1.81.1.tar.gz
cd leptonica-1.81.1
./configure
make
sudo make install

cd ~/src
wget https://github.com/tesseract-ocr/tesseract/archive/refs/tags/4.1.1.tar.gz -O tesseract-4.1.1.tar.gz
tar -zxvf tesseract-4.1.1.tar.gz
cd tesseract-4.1.1
./autogen.sh
export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig && export LIBLEPT_HEADERSDIR=/usr/local/include
./configure --with-extra-includes=/usr/local/include --with-extra-libraries=/usr/local/lib LDFLAGS="-L/usr/local/lib" CFLAGS="-I/usr/local/include"
make -j$(nproc)
sudo make install
sudo ldconfig
cd ~/src/openalpr/runtime_data/ocr/tessdata
wget https://github.com/tesseract-ocr/tessdata/raw/master/spa.traineddata

# compile OpenCV,  required by OpenALPR 
sudo yum -y install epel-release git gcc gcc-c++ cmake3 qt5-qtbase-devel python python-devel python-pip cmake python-devel python34-numpy gtk2-devel libpng-devel jasper-devel openexr-devel libwebp-devel libjpeg-turbo-devel libtiff-devel libdc1394-devel tbb-devel numpy eigen3-devel gstreamer-plugins-base-devel freeglut-devel mesa-libGL mesa-libGL-devel boost boost-thread boost-devel libv4l-devel

mkdir ~/src/opencv_build; cd ~/src/opencv_build
git clone https://github.com/opencv/opencv.git
git clone https://github.com/opencv/opencv_contrib.git
cd ~/src/opencv_build/opencv && mkdir build; cd build
cmake3 -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D INSTALL_C_EXAMPLES=ON -D INSTALL_PYTHON_EXAMPLES=ON -D OPENCV_GENERATE_PKGCONFIG=ON -D OPENCV_EXTRA_MODULES_PATH=~/src/opencv_build/opencv_contrib/modules -D BUILD_EXAMPLES=ON ..
# this is going to take some minutes...
make -j$(nproc)
sudo make install
sudo ln -s /usr/local/lib64/pkgconfig/opencv4.pc /usr/share/pkgconfig/  
sudo ldconfig
pkg-config --modversion opencv4

# finally, compile OpenALPR
cd ~/src/openalpr/src/build
cmake3 -DCMAKE_CXX_FLAGS="-std=c++11" -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
make
sudo make install

Finally as root user:

# if [ ! -f /etc/ld.so.conf.d/usrlocal.conf ] ;then echo "/usr/local/lib" > /etc/ld.so.conf.d/usrlocal.conf; echo "/usr/local/lib64" >>/etc/ld.so.conf.d/usrlocal.conf; fi
# sudo ldconfig -v

This is needed so alpr doesn't complain about finding libraries…

Test it:

# alpr --version

alpr  version: 2.3.0

# alpr --help
[...]

LONG VERSION AHEAD:

This guide assumes:

1. A sweet, bare-bones, really Minimal CentOS 7 installation, because I love it that way.
2. Commands will be issued by a non-root user that can sudo.

We'll need EPEL repo, wget, git and basic development tools like a C/C++ compiler etc.

$ sudo yum -y install epel-release
$ sudo yum -y install git wget
$ sudo yum groupinstall "Development tools"

Now we can install cmake (2) and cmake3 (3.17) which is found on EPEL repo

# sudo yum install cmake cmake3

Now the problem is cmake command is 2.8 and cmake3 the version we need. Thanks to this guide about Installing the latest git/cmake versions on RHEL/Centos.

The relevant part is to execute the following two commands, assuming you have installed cmake AND cmake3:

sudo alternatives --install /usr/local/bin/cmake cmake /usr/bin/cmake 10 \
--slave /usr/local/bin/ctest ctest /usr/bin/ctest \
--slave /usr/local/bin/cpack cpack /usr/bin/cpack \
--slave /usr/local/bin/ccmake ccmake /usr/bin/ccmake \
--family cmake

sudo alternatives --install /usr/local/bin/cmake cmake /usr/bin/cmake3 20 \
--slave /usr/local/bin/ctest ctest /usr/bin/ctest3 \
--slave /usr/local/bin/cpack cpack /usr/bin/cpack3 \
--slave /usr/local/bin/ccmake ccmake /usr/bin/ccmake3 \
--family cmake

Test cmake version:

$ cmake --version
cmake3 version 3.17.5

So in case we need to switch back cmake to be 2 or 3, we'll be able to change the default cmake version using:

$ sudo alternatives --config cmake

Prepare to build OpenALPR:

$ mkdir ~/src; cd ~/src 
$ git clone https://github.com/openalpr/openalpr.git
$ cd ~/src/openalpr/src && mkdir build; cd build

First shot, note that CMAKE_INSTALL_PREFIX has to be /usr/local in CentOS distro, otherwhise problems will arise regarding “Set runtime path” to a blank path for binaries and libraries when we execute the command sudo make install later.

$ cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
[...]
**ERROR: Required library Tesseract NOT FOUND.

If you get this other error: CMake 3.1 or higher is required. You are running version 2.8.12.2 then you didn't properly follow the step 1 of this guide, or just use cmake3 command instead of cmake.

So… ERROR: Required library Tesseract NOT FOUND.

Fine… Tesseract is a an OCR engine that OpenALPR needs so we need to install it too.

A) The Short path: we can install the default, Old tesseract version from our distro repos, which is 3.04 and it's aged.

$ sudo yum install tesseract tesseract-devel tesseract-langpack-spa 

Or we can go…

B) The Long path: install the shiny Tesseract version 4, reference: https://www.hoangdung.net/2020/01/how-to-install-tesseract-4-on-centos-7.html

First we need to compile Leptonica (a software for image processing and analysis), which is a dependency of Tesseract.

$ sudo yum -y install autoconf-archive libjpeg-turbo-devel libpng-devel libtiff-devel zlib-devel
$ cd ~/src
$ wget http://www.leptonica.org/source/leptonica-1.81.1.tar.gz
$ tar -zxvf leptonica-1.81.1.tar.gz
$ cd leptonica-1.81.1
$ ./configure
$ make
$ sudo make install

Check if libpng is installed: type whereis libpng and expect to see a directory; a blank line is not good

Check if leptonica is installed: type ls /usr/local/include and expect to see “leptonica”

Now let's compile the latest Tesseract released, 4.1.1 at the time of this writing. The process was smooth too:

$ cd ~/src
$ wget https://github.com/tesseract-ocr/tesseract/archive/refs/tags/4.1.1.tar.gz -O tesseract-4.1.1.tar.gz
$ tar -zxvf tesseract-4.1.1.tar.gz
$ cd tesseract-4.1.1
$ ./autogen.sh
$ export PKG_CONFIG_PATH=/usr/local/lib/pkgconfig && export LIBLEPT_HEADERSDIR=/usr/local/include
$ ./configure --with-extra-includes=/usr/local/include --with-extra-libraries=/usr/local/lib LDFLAGS="-L/usr/local/lib" CFLAGS="-I/usr/local/include"
$ make -j$(nproc)
$ sudo make install
$ sudo ldconfig

We have additional trained data here: https://github.com/tesseract-ocr/tessdata/

$ sudo yum install mlocate
$ sudo updatedb
$ locate traineddata
$ cd ~/src/openalpr/runtime_data/ocr/tessdata

Latest Spanish trained data:

$ wget https://github.com/tesseract-ocr/tessdata/raw/master/spa.traineddata

Let's try to configure openalpr using cmake again:

$ cd ~/src/openalpr/src/build

The next complain will be about OpenCV not installed.

$ cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
[...]
blablah OpenCV blah blah

So we need OpenCV, a real-time optimized Computer Vision library, another dependency of OpenALPR.

We have a “opencv-devel” package from official CentOS 7 repos, but…

$ sudo yum list opencv-devel

OpenCV from CentOS 7 repos is version 2.4.5 which won't be enough:

$ cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
[...]
ERROR: OpenCV version is not compatible : 2.4.5

Sweet… Let's cleanup if we installed it:

$ sudo yum remove opencv-devel opencv opencv-core

Why not, we're going to build OpenCV from sources to get OpenCV 4.x, which is a bit of extra pain, but this guide helped: https://linuxize.com/post/how-to-install-opencv-on-centos-7/

$ sudo yum -y install epel-release git gcc gcc-c++ cmake3 qt5-qtbase-devel python python-devel python-pip cmake python-devel python34-numpy gtk2-devel libpng-devel jasper-devel openexr-devel libwebp-devel libjpeg-turbo-devel libtiff-devel libdc1394-devel tbb-devel numpy eigen3-devel gstreamer-plugins-base-devel freeglut-devel mesa-libGL mesa-libGL-devel boost boost-thread boost-devel libv4l-devel

$ mkdir ~/src/opencv_build; cd ~/src/opencv_build

$ git clone https://github.com/opencv/opencv.git
$ git clone https://github.com/opencv/opencv_contrib.git

$ cd ~/src/opencv_build/opencv && mkdir build; cd build

$ cmake3 -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D INSTALL_C_EXAMPLES=ON -D INSTALL_PYTHON_EXAMPLES=ON -D OPENCV_GENERATE_PKGCONFIG=ON -D OPENCV_EXTRA_MODULES_PATH=~/src/opencv_build/opencv_contrib/modules -D BUILD_EXAMPLES=ON ..

This is going to take some minutes (nproc here is to inject number of CPUs available):

$ make -j$(nproc)

Everything ok? Then

$ sudo make install

$ sudo ln -s /usr/local/lib64/pkgconfig/opencv4.pc /usr/share/pkgconfig/  
$ sudo ldconfig

Check the OpenCV version available now:

$ pkg-config --modversion opencv4
4.5.3

Let's try again to configure OpenALPR

TLDR; sudo yum install curl curl-devel log4cplus-devel

$ cd ~/src/openalpr/src/build

$ cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
[...]
ERROR:   Could NOT find CURL (missing: CURL_LIBRARY CURL_INCLUDE_DIR)

Great… let's install curl-devel

$ sudo yum install curl curl-devel

Trying again…

$ cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
[...]
ERROR: Required library log4cplus NOT FOUND.

Fine, let's install log4cplus-devel

$ sudo yum install log4cplus-devel

Should be OK now!

$ cmake -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
[...]
-- Build files have been written to: .../src/openalpr/src/build

Let's finally compile this thing.

$ cd ~/src/openalpr/src/build
$ cmake -DCMAKE_CXX_FLAGS="-std=c++11" -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..
$ make
$ sudo make install

Long version:

$ cd ~/src/openalpr/src/build

Let's try make:

$ make
[...] error: #error "OpenCV 4.x+ requires enabled C++11 support"
[...]
make[2]: *** [video/CMakeFiles/video.dir/videobuffer.cpp.o] Error 1

It's complaining about enabling C++11 support which we can fix using -DCMAKE_CXX_FLAGS="-std=c++11" at cmake call (took me a bit of research to fix):

Cleanup first:

$ make clean

Redo the cmake call:

$ cmake -DCMAKE_CXX_FLAGS="-std=c++11" -DCMAKE_INSTALL_PREFIX:PATH=/usr/local -DCMAKE_INSTALL_SYSCONFDIR:PATH=/etc ..

Try again make and it should go well now:

$ make

$ sudo make install

Evertyhing ok?

Let's try if the alpr command is finally available:

$ alpr
alpr: error while loading shared libraries: libopenalpr.so.2: cannot open shared object file: No such file or directory

Sweet… Another clusterf*ck I had to research a bit.

Create a new file named /etc/ld.so.conf.d/usrlocal.conf and put these lines:

/usr/local/lib
/usr/local/lib64

$ sudo ldconfig -v

Let's try again:

$ alpr --version
alpr  version: 2.3.0

Get an image of a vehicle plate, say plate.jpg.

$ alpr --help

Try an image (by default “us” country is used).

$ alpr plate.jpg

You can try “eu” as country and limit guess at 2 top matches:

$ alpr -c eu -n 2 plate.jpg
plate0: 2 results
  - 2008ZGZ    confidence: 94.4568
  - 20Q8ZGZ    confidence: 87.0557

~~DISCUSSION|Comentarios~~